Fluoroquinolone Residues in Raw Cow's Milk

Czech J. Food Sci. Vol. 29, 2011, No. 6: 641–646

Fluoroquinolone Residues in Raw Cow’s Milk

Pavlína Navr átilová, Ivana Borkovcová, Jana Vyhnálková and Lenka Vorlová

Department of Milk Hygiene and Technology, Faculty of Veterinary Hygiene and Ecology, University of Veterinary and Pharmaceutical Sciences Brno, Brno, Czech Republic

Abstract

Navrátilová P., Borkovcová I., Vyhnálková J., Vorlová L. (2011): Fluoroquinolone residues in raw cow’s milk. Czech J. Food Sci., 29: 641–646.

The aim of the study was to monitore the levels of fluoroquinolone residues in bulk samples of raw cow’s milk. The bulk samples of raw cow’s milk (n = 150) were obtained from 58 different milk suppliers in the South Moravian and Vysočina Regions. The samples were analysed by reverse phase high-performance liquid chromatography method with fluorescence detection and gradient elution. 87.3% of the raw milk samples were positive for the fluoroquinolones residues. Flumequine was present in none of the samples. The levels of other fluoroquinolones investigated were below the recommended maximum residue limit. The results of the study indicate that fluoroquinolones are frequently administered to the dairy cows in spite of the recommendations of CVMP. The most frequently determined was enrofloxacin and its indicator residue, i.e. ciprofloxacin. An efficient control of the veterinary drugs residues in milk is very important to ensure the safety of the milk and milk products.

Keywords: veterinaty drug; RP HPCL; food safety; MRL; LOD; LOQ

The monitoring of the veterinary drugs residues bacterial agents in both human and veterinary is an important part of the food safety control in medicine. Fluoroquinolones are effective for the raw materials and foods of animal origin. To ensure therapy of serious diseases, e.g. septicaemia, gas- the safety of food of animal origin for consumers, troenteritis, and respiratory diseases caused by maximum residue limits (MRL) of veterinary drugs Gram-negative bacteria. They are also used for used with the food animals have been set for raw the treatment of infections of the urinary tract materials of animal origin (Botsoglou & Fle- and skin, and of infections of soft tissues caused touris 2001; Commission Regulation 37/2010). by Gram-negative and certain Gram-positive aero- Since 1976, when the first monofluoroquinolone bic bacteria. They are effective in the therapy for flumequine was developed, very many fluoroqui- mycoplasma infections and infections caused by nolone representatives have been synthetised and atypical bacteria. Incidence of side adverse effects described. Compared with the classic 1st generation of fluoroquinolone treatment, particularly in hu- quinolones that were characterised by a narrow man medicine, has been reported (Flomenbaum spectrum of action and poor tissue penetration et al. 2006; CVMP 2007). ability, fluoroquinolones feature many attributes In veterinary medicine, they are useful especially of ideal antimicrobials (Sarkőzy 2001; Emmer- in the therapy for gastrointestinal and respiration son & Jones 2003). Fluoroquinolones are used infections (Botsoglou & Fletouris 2001). Fluo- for the treatment of infections caused by various roquinolones are used in intramammary prepara-

Supported by the Ministry of Education, Youth and Sports of the Czech Republic, Project No. MSM 6215712402.

641 Vol. 29, 2011, No. 6: 641–646 Czech J. Food Sci. tions for the treatment of mastitis in lactating dairy enrofloxacin (ENRO) and its indicator residue = cows, for dairy cows during dry periods, and also sum of enrofloxacin and ciprofloxacin (CIPRO), for mastitis prevention (Gruet et al. 2001). and the residues of marbofloxacin (MARBO), Although fluoroquinolones found a wide applica- danofloxacin (DANO) and flumequine (FLU), and tion in primary agriculture, their administration to to determine their concentrations. food animals has been discussed in recent years in connection with the increased incidence of resistance among human pathogenic micro-organisms. Material and Methods From the onset of the bacterial resistance point of view, fluoroquinolones belong to the high-risk Milk samples. The bulk samples of raw cow’s groups of antimicrobial drugs (WHO 1998; CVMP milk (n = 150) were obtained from 58 different 2007). The World Health Organisation (WHO) milk suppliers in the South Moravian and Vysočina and Word Organisation of Animal Health (OIE) Regions. The samples of bulk milk were collected have defined quinolones as “critically important from the milk collection route of the dairy plant. antimicrobials” for human and animal health, The milk samples were kept at temperatures not respectively (FAO 2007). exceeding 4°C until the testing time. Due to extra-label use of veterinary drugs or Chemicals and materials. The standards of fluo- noncompliance withdrawal periods, much higher roquinolones, i.e. of enrofloxacin (17849), flume- residue levels might appear in the edible animal quine (F7016), marbofloxacin (34039), danofloxacin products (Botsoglou & Fletouris 2001). The (33700), ciprofloxacin (17850), and norfloxacin published data indicated that violative residues (N 9890) were purchased from Sigma Aldrich (St. were found in raw or heat treated milk samples in Louis, USA). The chemicals for HPLC analysis, i.e. countries all over the world (Botsoglou & Fle- acetonitril, methanol and dichloromethane (HPLC touris 2001; Tolentino et al. 2005; Fonseca et grade) were obtained from Merck (Darmstadt, al. 2009; Kaya & Filazi 2010; Addo et al. 2011). Germany). Trifluoroacetic acid and phosphoric Quality control of the purchased raw cow’s milk acid (suprapure grade) were from Fluka Chemie implies that 0.16% samples showed violative con- AG (Buchs, Switzerland) and Sigma Aldrich (St. centrations of residues (Kopunetz 2010). Louis, USA). NaOH (analytical grade), was pur- Within the EU, each state is obliged to monitor chased from Penta (Prague, Czech Republic). The foodstuffs for the residues of veterinary drugs hardware included an analytical balance (Kern & and to present a National Residue Monitoring Sohn GmbH, Balingen, Germany), a cooling cen- Plan that takes into account the specific situa- trifuge (Mechanika Precyzyjna, Bytom, Poland), tion in its country (Botsoglou & Fletouris a rotary vacuum evaporator (Bűchi, Postfach, 2001). In the Czech Republic, milk is monitored Switzerland), and a vortex (Merck, Darmstadt, for the presence of veterinary drugs residues in Germany). accredited laboratories as part of their food chain Preparation of standard fluoroquinolone so- monitoring for exogenous substances, and also lutions. The stock solutions at active substance during quality checks of the raw milk. Screening concentration of c = 1 g/l were first prepared by tests for veterinary drugs residues in milk rely dissolving an adequate amount of the chemothera- mainly on rapid tests and microbial inhibition peutic in 1 ml of 0.1 mol/l NaOH and adding deion- methods (plate diffusion method) (Botsoglou ised water to a total volume of 25 ml. The stock & Fletouris 2001; Czech National Reference solutions of chemotherapeutics were then further Laboratory 2008). In targeted testing for fluoroqui- diluted with deionised water to obtain working nolone residues, modern analytical methods such solutions with a concentration in the 0.1–1.0 mg/l as the high-performance liquid chromatography range that were then used to prepare fortified milk (HPLC) analysis and mass spectrometry are used samples for the method validation. (Hernández-Arteseros et al. 2002; Marazuela Preparation of samples for HPLC analysis. & Moreno-Bondi 2004). Quinolones were extracted from the milk sam- The objective of the present study was to give the ples by the liquid/liquid extraction method. Con- information on the presence of fluoroquinolone centrated phosphoric acid solution (0.5 ml) was residues for which maximum residue limits were added to 5 ml of milk and vortexed for 5 seconds. set in raw cow’s milk. These imply the residues of Acetonitrile (10 ml) was pipetted into the mixture

642 Czech J. Food Sci. Vol. 29, 2011, No. 6: 641–646 for extraction by vortexing for 1 minute. The re- marbofloxacin, 10.0–400.0 μg/l for norfloxacin, sulting mixture was then centrifuged at 4000 rpm 6.5–260 μg/l for ciprofloxacin, 2.0–80.0 μg/l for for 10 min at 5°C. The supernatant was collected danofloxacin, 6.2–250.0 μg/l for enrofloxacin, and (10 ml) and acetonitrile was evaporated. Two ml 1.2–48.0 μg/l for flumequine. of dichloromethane were added to the residue, The limit of detection (LOD) and the limit of and the mixture was again vortexed for 10 s and quantification (LOQ) were determined from the centrifuged at 4000 rpm for 10 min at 5°C. The analysis of the matrix samples fortified with fluo- 1 ml sample collected from the upper aqueous roquinolone standards, and evaluated as 3 signal layer was diluted with the mobile A phase at a to noise ratio (S/N) and 10 S/N in µg/kg. The re- ratio of 1:1 for HPLC analysis. peatability and recovery were determined from 12 HPLC conditions. The HPLC assay was per- parallel analyses of the milk samples fortified with formed on an Alliance 2996 liquid chromatograph the standards solutions of known concentrations, with a 2475 fluorescence detector (Waters, Mil- 0.5, 1.0, and 1.5 MRL. To determine the analyte ford, USA). The separation was performed on recovery for the individual fluoroquinolones, the an Atlantis T3 chromatographic column 4.6 × analytical results obtained with the fortified sam- 250 mm with the particle size of 5 μm (Waters, ples were compared with an externally generated Milford, USA). The mobile A phase consisted of calibration curve of the standards. Their mean 0.2% trifluoroacetic acid, the mobile B phase of a values are summarised in Table 1 together with methanol + acetonitrile mixture (1:5), the gradi- the coefficients of variation for repeatability. The ent was linear from 25% B phase to 80% B phase, decision limit CCα and the detection capability and the flow rate was 0.7 ml/min. The injection CCβ as defined by the Commission Decision 657 volume was 20 μl and the column temperature (2002) were determined by analysing the matrix was 30°C. The wavelengths selected for the fluo- milk samples at the concentration levels of 0.5, rescence detection were λex/λem = 280/450 nm for 1.0, and 1.5 MRL of each analyte. MARBO, CIPRO, DANO, and ENRO, and λex/λem = Statistical evaluation. Basic statistical evalua- 312/366 nm for FLU. tion of the results was performed using Statistica Method validation. An external standard method 7 statistical software (StatSoft, Prague, Czech was used for both qualitative and quantitative Republic). evaluations. Each sample was measured in two parallel assays, and at the same time, a blank sample per series was also measured. Norfloxacin was Results and Discussion used as a surrogate standard. The milk samples not containing any of the analytes monitored were Food safety is an important issue in the EU. analysed to test for the method selectivity. Foodstuffs that contain violative antibiotic resi- The calibration curves of the matrix samples dues constitute public health hazards including were generated in the range of 3.5–140.0 μg/l for toxicological (direct toxic effect), microbiological

Table 1. Validation results of HPLC with fluorescence detection: recovery, repeatability, LOD, LOQ, R2, CCα and CCβ values

Parameter Marbofloxacin Ciprofloxacin Danofloxacin Flumequine Enrofloxacin Recovery (%) 89.1 97.1 91.6 82.6 103.0 RSD (%) 5.4 10.1 3.4 3.7 7.1 R2 0.98 0.99 0.99 0.99 0.98 LOD (µg/kg) 7.0 6.0 4.0 3.0 6.0 LOQ (µg/kg) 21.0 18.0 12.0 10.0 18.0 CCα (µg/kg) 79 108 31 51 104 CCβ (µg/kg) 83 116 32 52 108

RSD = relative standard deviation; LOD – the limit of detection; LOQ – the limit of quantification;n – number of measure- ments; R2 – linearity; CCα - decision limit; CCβ – detection capability

643 Vol. 29, 2011, No. 6: 641–646 Czech J. Food Sci.

(transmittance of antibiotic resistance, adverse food animals is prohibited (Commission Regula- effects on the ecology of human intestinal micro- tion 1662/2006). MRL levels set by EU legislation flora), and immunopathological (hypersensitivity for fluoroquinolones in milk differ widely between reactions) hazards (Roeder & Roeder 2000). The the individual compounds, from 0.03 mg/kg for incidence of some side adverse effects of fluoro- danofloxacin and 0.05 mg/kg for flumequine to quinolone treatment has been described. The most 0.075 mg/kg for marbofloxacin and 0.1 mg/kg for important side effects of fluoroquinolones include the sum of enrofloxacin and ciprofloxacin residues renal impairment, ocular disorders, and damage (Commission Regulation 37/2010). to articular cartilage. Articular cartilage damage For the detection of the selected fluoroquinolones is irreversible and for that reason the preparations in raw cow’s milk, HPLC analysis with fluorescence containing fluoroquinolones must not be adminis- detection was validated (Table 1). In the process tered to adolescents and pregnant women. Other of the method validation, CCα and CCβ were adverse effects include, e.g., frequent gastrointes- determined in compliance with the legislation tinal disturbances (diminished appetite, nausea, (Commission Decision 657/2002) (Table 1). abdominal pain, vomiting, diarrhoea), headache, Chung et al. (2009) conducted a study in which insomnia, tension, or skin allergy manifestations they used HPLC analysis with fluorescence detec- (Flomenbaum et al. 2006). tion to determine ENRO and CIPRO residues. The Committee for Medicinal Products for Vet- Compared with our study, their analysis validation erinary Use (CVMP) classifies quinolone prepa- recovery was higher for CIPRO (101.3%) but lower rations into critically important antimicrobials for ENRO (72.4%). The limits of detection were that should be administered in accordance with in their study 6.5 μg/kg and 1.1 μg/kg, and the prudent use principles. Prudent use of quinolones limits of quantification 23.2 μg/kg, and 3.0 μg/kg is defined as the practices that maximise thera- for CIPRO and ENRO, respectively. Hermo et al. peutic effect while minimising the emergence of (2008) developed and applied various methods of resistance (CVMP 2007; WHO 1998). At present, liquid chromatography for the determination of injection preparations containing danofloxacin, quinolone residues in milk at MRL levels, and com- enrofloxacin, difloxacin, and marbofloxacin are pared them with the mass spectrometry method. registered in the Czech Republic for use in lac- For all quinolones, a recovery in excess of 80% in tating dairy cows. Raw cow’s milk, irrespective of the validation of LC methods was attained. The whether it is to be processed or sold directly to LC-FD method produced LOD in the 3–8 ng/g consumers, must meet the quality standards set by range. The limits of quantification (LOQ) were EU legislation. Neither raw cow’s milk nor milk of determined at levels below the MRL with the only other animal species may be placed on the market exception of MARBO in LC-UV. In compliance if it contains the residues of pharmacologically with EU legislation, they determined CCα and CCβ active substances in quantities in excess of the set values of the individual methods when performing MRLs, or if it contains substances whose use in the validation. Their LC-FD method validation of

Table 2. Concentrations of residues of fluoroquinolones determined in bulk samples of raw cow’s milk

Fluoroquinolone (µg/l) sum of enrofloxacin enrofloxacin ciprofloxacin marbofloxacin danofloxacin and ciprofloxacin n1/n 115/150 122/150 131/150 22/150 33/150 x–(µg/l) 10.24 8.23 16.70 5.42 3.41 Min (µg/l) 1.70 2.70 2.70 5.00 2.00 Max (µg/l) 18.60 17.40 27.40 6.00 10.20 SD (µg/l) 1.64 2.72 5.03 0.28 1.84 CV (%) 16.01 33.08 30.09 5.24 54.00

– n1 – number of positive samples; n – total number of examined samples; x - mean; min and max – minimum and maximum values; SD – standard deviation; CV – coefficient of variation

644 Czech J. Food Sci. Vol. 29, 2011, No. 6: 641–646 the fluoroquinolones analysed gave lower CCβ recommend the use of 18% DANO administered and CCα values that those in the present study subcutaneously for the treatment of mastitis and (Table 1), with the exception of CCα value for other infectious diseases in dairy cattle provided MARBO, which is the same as in our study. that a 3-day withdrawal period is observed. The residues of ENRO and of its main metabolite The HPLC method failed to demonstrate the pres- CIPRO were detected in 87.3% of the total number ence of flumequine in the bulk samples of raw cow’s of the samples tested (Table 2). The presence of milk. The results are consistent with the fact that only ENRO was determined in 10 samples, and the no flumequine-containing drug for use in lactating presence of only CIPRO, its main metabolite, was dairy cows is registered in the Czech Republic. determined in 13 samples. The maximum values of enrofloxacin and ciprofloxacin residues demonstrate that their levels in the samples were very Conclusions low (Table 2). The sum of enrofloxacin residues and its indicator residues (= sum of enrofloxacin HPLC with fluorescence detection was used to and ciprofloxacin) did not exceed the MLR value detect the residues of the monitored fluoroqui- (0.1 mg/kg) in any of the bulk sample. In 13% of nolones (ENRO, CIPRO, MARBO, and DANO) the samples, the total concentration of ENRO and in bulk samples of raw cow’s milk. The results CIPRO residues was less than 10 ± 5.03 μg/l, in of the study indicate that fluoroquinolones are 65.6% of the samples it was 10–20 ± 5.03 μg/l, and frequently administered to dairy cows in spite in 21.4% it was in the 20–25 ± 5.03 μg/l range. Only of the recommendations of CVMP. The residue 4 samples (3%) revealed concentrations higher levels determined did not exceed MRL in any of than 25 ± 5.03 μg/l. The highest ENRO and CIPRO the samples. Flumequine was not present in any concentration found was 27.4 ± 5.03 μg/l. The of the samples. The most frequently determined total sum of ENRO residues eliminated from the in the samples was enrofloxacin and its indicator body is made up, besides the original substance, residue, i.e. ciprofloxacin. An efficient control of of two main metabolites, i.e. enrofloxacin amide the residues in milk is very important to ensure and ciprofloxacin, and, to a lesser degree, other the safety of milk and milk products. metabolites. If enrofloxacin is administered intra- venously to dairy cows, ciprofloxacin may continue to be detected at relatively high concentrations References over a longer period than enrofloxacin itself (Bot- soglou & Fletouris 2001). Addo K.K., Mensah G.I., Aning K.G., Nartey N., Nipah MARBO residues were detected in only 14.6% G.K., Bonsu, C., Akyeh M.L., Smits H.L. (2011): Micro- of the total of the samples examined (Table 2). biological quality and antibiotic residues in informally MARBO is a fluoroquinolone that is used to treat marketed raw cow milk within the coastal savannah zone bovine respiratory diseases. The studies moni- of Ghana. Tropical Medicine & International Health, 16: toring the elimination of MARBO residues in 227–232. livestock demonstrated that 73–89% of the total Aliabadi F.S., Landoni M.F., Lees P. (2003): Pharma- sum of residues in lactating dairy cows milk is cokinetics (PK), pharmacodynamics (PD), and PK-PD made up of the original substance (Botsoglou integration of danofloxacin in sheep biological fluids. & Fletouris 2001). Antimicrobial Agents and Chemotherapy, 47: 626–635. DANO residues were present in only 22% of Botsoglou N.A., Fletouris D.J. (2001): Drug Residues samples. The residue concentrations were in the in Foods. Marcel Dekker, New York. 2.7–27.4 ± 5.03 µg/l range (Table 2). DANO exhibits Chung H.H., Lee J.B., Chung Y.H., Lee K.G. (2009): Analy- a wide spectrum of activity – it is effective against sis of sulfonamide and quinolone antibiotic residues in both Gram-negative and Gram-positive bacteria Korean milk using microbial assays and high performance and mycoplasma, but it is less effective against liquid chromatography. Food Chemistry, 113: 297–301. anaerobic microorganisms (Aliabadi et al. 2003). Commission Decision No. 657/2002 of 12 August 2002 The transfer of DANO from blood to milk was very implementing Council Directive 96/23/EC concerning the rapid with maximum milk concentration having performance of analytical methods and the interpretation been reached only 8 h after the administration. On of results. Official Journal of the European Communities, the basis of their study, Mestorino et al. (2009) L 221: 8–36.

645 Vol. 29, 2011, No. 6: 641–646 Czech J. Food Sci.

Commission Regulation (EU) No. 37/2010 of 22 December LC-MS/MS and validation in line with regulation 2009 on pharmacologically active substances and their 2002/657/EC. Analytica Chimica Acta, 613: 98–107. classification regarding maximum residue limits in food- Hernández-Arteseros J.A., Barbosa J., Compañó R., stuffs of animal origin. Official Journal of the European Prat M.D. (2002): Analysis of quinolone residues in Communities, L 15: 1–76. edible animal products. Journal of Chromatography A, Commission Regulation (EC) No. 1662/2006 of 6 November 945: 1–24. 2006 amending Regulation (EC) No. 853/2004 of the Euro- Kaya S.E, Filazi A. (2010): Determination of antibiotic pean Parliament and of the Council laying down specific residues in milk samples. Kafkas Universitesi Veteriner hygiene rules for food of animal origin. Official Journal of Fakultesi Dergisi, 16: S31–S35. the European Communities, L 320: 1–10. Kopunetz P. (2010): Přehledy jakosti nakupovaného mléka CVMP (2007): Public statement on the use of (fuoro) qui- v roce 2010. ČMSCH, Hradištko: 38. Available at http:// nolones in food-producing animals in the European Un- www.cmsch.cz/laboratore-pro-rozbor-mleka-lrm/ ion: development of resistance and impact on human Marazuela M.D., Moreno-Bondi M.C. (2004): Mul- and animal health. EMEA/CVMP/SAGAM, London: tiresidue determination of fluoroquinolones in milk by 1–18. Available at http://www.ema.europa.eu/docs/ column liquid chromatography with fluorescence and en_GB/document_library/Public_statement/2009/10/ ultraviolet absorbance detection. Journal of Chromatog- WC500005152.pdf raphy A, 1034: 25–32. Emmerson A.M., Jones A.M. (2003): The quinolones: Mestorino N., Marchetti M.L., Turic E., Pesoa J., Er- decades of development and use. Journal of Antimicrobial recalde J. (2009): Concentrations of danofloxacin 18% Chemotherapy, 51: 13–20. solution in plasma, milk and tissues after subcutaneous FAO (2007): Joint FAO/WHO/OIE Expert Meeting on injection in dairy cows. Analytica Chemica Acta, 637: Critically Important Antimicrobials. Report of a meeting. 433–439. FAO/WHO, Rome and Geneva: 1–18. Available at http:// Roeder R.A., Roeder M. (2000): Antibiotics in food of www.who.int/foodborne_disease/resources/Report%20 animal origin. In: Francis J.F. (ed.): Food Science and joint%20CIA%20Meeting.pdf Technology. John Wiley & Sons, Inc., New York: 54–63. Flomenbaum N.E., Goldfrank L.R., Hoffman R.S., Sarkőzy G. (2001): Quinolones: a class of antimicrobial Howland M.A., Lewin N.A., Nelson L.S. (2006): Gold- agents. Veterinarni Medicina, 46: 257–274. frank’s Toxicologic Emergencies. 8th Ed. McGraw-Hill, The Czech National Reference Laboratory (2008): Methodo- New York. logical guidelines for the detection residues of inhibitory Fonseca G.P., Cruz A.G., Faria J.A.F., Silva R., Moura substances in tissues, milk, eggs and foods of animal M.R.L., Carvalho L.M.J. (2009): Antibiotic residues in origin. State Veterinary Institute, Jihlava: 14. Brazilian UHT milk: a screening study. Ciencia e Tecno- Tolentino R.G., Perez M.N., Gonzalez G.D., Leon logia de Alimentos, 29: 451–453. S.V., Lopez M.G., Flores G.P. (2005): Determination Gruet P., Maincent P., Berthelot X., Kaltsatos V. of the presence of 10 antimicrobial residues in Mexican (2001): Bovine mastitis and intramammary drug deliv- pasteurized milk. Interciencia, 30: 291–294. ery: review and perspectives. Advanced Drug Delivery WHO (1998): Use of quinolones in food animals and poten- Reviews, 50: 245–259. tial impact on human health. Geneva: Report of a WHO Hermo M.P., Nemutlu E., Kir S., Barrón D., Barbosa Meeting: 1–18. Available at http://www.who.int/emc J. (2008): Improved determination of quinolones in milk Received for publication January 13, 2011 at their MRL levels using LC-UV, LC-FD, LC-MS and Accepted after correctiopns April 18, 2011

Corresponding author: MVDr. Pavlína Navrátilová, Ph.D., Veterinární a farmaceutická univerzita Brno, Ústav hygieny a technologie mléka, Palackého 1/3, 612 42 Brno, Česká republika tel.: + 420 541 562 716, e-mail: [email protected]

646